Fifty years ago, Heyerdahl and Skjolsvold (1956, 1990) collected material from five archaeological sites in the Galapagos Islands. They retained earthenwares of possible precolumbian origin and discarded ceramic, metal, and glass artifacts postdating the arrival of the Spanish in A.D. 1535. Consequently, they argued that each site was formed as the results of a series of discard events from unrelated short-term occupations extending from the precolumbian to the historical era, and that the earthenwares represented occasional visits by fishermen from precolumbian Peru and Ecuador. In 2005, we re-excavated the sites and collected all the excavated materials. Our results show that each class of material, irrespective of age or origin, was distributed spatially and stratigraphically in the same pattern, contradicting the former assumption of multiple, unrelated occupations. We reject the palimpsest model in favor of the null hypothesis of single-phase site occupation. Analysis of putatively precolumbian pottery using optically-stimulated luminescence dating indicates that it is mostly of historical age. Radiocarbon dating confirms that the archaeological sites are younger than the sixteenth century. Research on sedimentary cores shows probable anthropogenic impacts as restricted to the last 500 years. We conclude that there was no human occupation in the Galapagos Islands until the historical era.